Page 1 of 1

$\chi^2$ Value Question - COSMOMC

Posted: April 14 2020
by Lawrence Kazantzidis
Hi everyone,

I am currently working on a project using cosmomc and I have successfully performed some runs including CMB (TT+lowTEB+lensing), BAORSD, and DES likelihoods and obtained [math]. I would like to ask a few questions regarding the obtained $\chi^2$ value.

1) Usually $\chi^2$ is normalized by dividing it with the degree of freedoms(dof). What is the dof of the CMB likelihoods (TT+lowTEB+lensing)
in order to have a qualitative hint of my fit?

2) I am not quite sure why there is an errorbar in the total $\chi^2$ value. Do I have to also include the value of the prior?

Re: $\chi^2$ Value Question - COSMOMC

Posted: April 14 2020
by Antony Lewis
I think the normalization for lowTEB is fairly arbitrarily. It's clearer with the 2018 likelihoods.

The chi2 values you quote are presumably the mean and standard deviation of the chi2 for each point in the sampled parameter space. You can do a minimization run if you actually want a best-fit value.